Value and reliability of DNAPL investigation programs

An analytical method for estimating the value of investigation for sites contaminated with Dense, Non-Aqueous Phase Liquids (DNAPLs) is developed in this dissertation. DNAPL contamination is difficult to locate and characterize. No practical methods exist for quantifying variability in DNAPL contamination or for estimating the value of investigating DNAPL-contaminated sites. The first task performed in this research is to develop a decision framework for DNAPL-contaminated sites so that the value of investigation may be studied. Several additional tasks focus on overcoming current difficulties to the practical application of the theoretical framework. The second task is to quantify variability in the spatial distribution of DNAPL for a scaled, simplified representation of several case-study sites. This task is intended to develop a starting point for estimating the value of investigation. The results indicate that there is substantial variability in the bounds of DNAPL contamination. The scaled volume of release is the dominant factor affecting the distribution of DNAPL. In addition, the presence or absence of a low-permeability layer beneath the DNAPL-release location plays a significant role in the migration of DNAPL. The third task is to quantify the reliability of investigation effort in order to model the probability that a program will find DNAPL. The models indicate that substantial or even unreasonable investigation effort may be required to locate DNAPL with current invasive technologies. The fourth task is to develop an analytical method for revising uncertainty in the geometry of DNAPL contamination, given any outcome from a one-dimensional investigation program. The approximations agree well with simulated and exact solutions, when available. The final task is to incorporate the results of all previous tasks to study the value of investigation for a case-study site. Examples demonstrate that the investigation objective of locating DNAPL is difficult to justify on an, expected-cost basis because the cost of the effort can easily exceed its value. An alternative objective is to reduce uncertainty in the geometry of DNAPL contamination. Examples show that the cost to achieve this objective is much less than for the location objective. The value of this type of investigation is shown to depend on the amount of effort and the difference in expected costs between the decision alternatives. There may be value to conducting a two-stage investigation for conditions that do not warrant a single-stage investigation. An example also illustrates that the value of investigation may decrease with additional investigation effort for certain conditions.